156 
2.2 The choice of the benchmarks 
3. MEASUREMENTS AND PRECISION 
We have then gone on with on-the-spot inspections to detect the 
existing benchmarks, to verify their conditions and asses where 
it was necessary to integrate them. 
In some cases, in proximity of the old benchmarks showing 
difficulty in the access and in the measure, we have introduced 
one of a new kind (vertical axis nail or horizontal axis bolt) with 
the aim of strengthening the signalisation. 
A part of all benchmarks still in existence, belonging to the old 
levelling lines and having been found, has been included into 
the actual network, while as to the other ones we have worked 
out their heights either as the passage point or through direct 
measure from the nearest levelling station. 
On the whole the network consists of n°192 benchmarks, with 
an average density of one each 500 m more or less, and in 
particular of 7 points/km 2 in the area nearer the CNR complex 
(for a surface of 15 km 2 ) and of 2 points/km 2 in the outer parts 
(for a surface of 45 km 2 ). 
According to their source we can count: 
- n° 37 benchmarks set up by I.G.M.; 
n° 30 benchmarks set up by I.T.G. (University of Pisa, 
Istituto di Topografia e Geodesia); 
- n° 92 benchmarks set up by the C.N.R; 
n° 33 benchmarks set up by different Institutes and Bodies. 
23 The choice of the reference benchmarks 
As far as the choice of the reference benchmark is concerned - 
never forgetting that the gradients are nonetheless relative 
determinations and that the subsidence measurements are 
nothing absolute in themselves - we have established two 
separate points, distant from each other, in areas which are 
presumably steady and not much affected by well boring and 
water production/reinjection operations (always taking into 
account the geological appraisals coming from a centuries-old 
experience in the survey of the area to be examined): 
- the benchmark of Migliarino Pisano (IGM 31/8 Chiesa 
Cappella Borghese Salviati), originally taken on as a reference 
point for the vertical movements of the plain of Pisa, which 
belongs to the basic national network of the I.G.M. of 1886 and 
1951; 
- the benchmark of S. Giuliano Terme (IGM 15 - III - C 1920 
horizontal spa building) was taken on as a reference point and 
connected to that of Migliarino Pisano. 
From "The vertical movements of the ground in the plain of 
Pisa since 1920, inferred from the comparison of levellings". B. 
Palla, T. Cetti, M. Poggianti, E. Mengali, A. Bartolini, 
Provincia e Comune di Pisa, April 1976: 
" The choice of the benchmark I.G.M. 1951 of the Chiesa 
Cappella Borghese Salviati in Migliarino Pisano, for which we 
have taken on the absolute height (at the medium sea level in 
Genoa 1942) of +2.8762 m fixing the situation in 1951, 
depended on the following factors: 
the benchmark measured in 1951 belonged also to the basic 
I.G.M. network of 1886 as n° 11 of the line n° 42 Pisa - Sarzana 
with reference to the Genoa medium sea level of 1897; 
it lies on an area which is very quiet from the point of view of 
the traffic both on the railway and on the road, and unchanged 
by town plans since long before 1951; 
it is located on a longstanding handmade building (according to 
the gathered information it dates back to 1795) 
the height difference between that and the n° 11-C of Casa 
dellTmmaginetta, around 3 km. far in the north direction and 
belonging to the same line Pisa - Sarzana, gave the result of 2,6 
mm. in absolute value over a time of more than 20 years... 
The first measurement (the zero measure for the future) began 
in 1998 from the benchmark I.G.M. n°136 (Pisa - La Fontina) 
during the month of October and it is scheduled to be completed 
within March 1999. 
Such a benchmark consists of a bolt driven inside at the 
beginning of the Madiceo Aqueduct, in Pratale Street, on the 
Calcesana way side, a few metres away from the CNR complex 
and it forms the starting point of the polygon 3, which has been 
run across clockwise, like all the following polygons. 
The instrument that has been used is the automatic level Wild 
NA 2, provided with parallel plate. Since in the levelling of the 
entire network only one staff has been adopted, instead of two 
like sometimes it is possible to do in order to reduce the work 
time, we did not have to check out the calibration, that is the 
determination of the distance from the origin of the graduation 
of the heel for each of them, so as to take it into account in the 
following calculations. 
Although the use of one stadia rot only, as everybody knows, 
makes the measurement operations a bit longer, it reduces 
considerably the risk for blunders in the subsequent 
calculations, which are extremely improbable but still possible 
at least in theory, and it avoids the negative influence of 
accidents able to invalidate the steadiness of the heel, if any. 
As mentioned before, every line segment included between two 
benchmarks, which are on average about 500 metres distant 
from each other, has been levelled twice, by different operators 
and on the opposite way in order to achieve independent results. 
As to the precision of the determinations of the gradients 
between the benchmarks and therefore their absolute heights 
(referred to S. Giuliano Terme point), we have made use of 
technical rules and operative procedures typical of the precision 
levelling, but considering the special interest in the knowledge 
of the steadiness of the territory studied, even from a scientifical 
standpoint, we have reduced the tolerance for the closing 
spreads on each line lower than 3 mm. 
It has proved necessary to repeat the levelling only twice in the 
stretches that have been measured so far, since the discordance 
between the outward gradient and the return one exceeded the 
value we had assessed as the one admittable. 
4. THE SOFTWARE FOR THE FIRST ELABORATION 
OF THE MEASURES (LIVEWIN 2.0). 
The use of computer procedures for the processing of the 
measures is an important step in order to enhance productivity 
and quality. In fact, it allows to accelerate the process, the 
possibility of data transfer, to elaborate and in general to 
manage information with no phases of transcription (potential 
causes of errors are the more probable the more complex the 
process is) and it reveals itself to be fundamental. Besides, the 
computer elaboration lets us bring forth the calculation with 
decimal approximations only on the final result. 
For the levelling above-mentioned, we thought it right to 
develop a special software able to execute the preliminary 
analysis and prepare data for the following calculations. 
The procedure has been developed through Microsoft 
VisualBasic that allows the realization of a 32 bit application for 
Windows 95/98 environment (a friendly user interface). 
Going ahead with the job, all the functionalities of the software 
were tested in order to eliminate the programming bugs while 
making it suitable for the most frequent demands. It is a very 
flexible product, able to fit different situations of use. 
The start window of LiveWin 2.0 provides the possibility to